Degradation mechanism analysis and State-of-Health estimation for lithium-ion batteries based on distribution of relaxation times. (1st November 2022)

• Record Type: Journal Article
• Authors: Zhang, Qi; Wang, Dafang; Schaltz, Erik; Stroe, Daniel-Ioan; Gismero, Alejandro; Yang, Bowen
• Is Part Of: Journal of energy storage. Volume 55:Part A(2022)
• Publication Date: 2022-11-01
• Subjects: Lithium-ion battery -- Distribution of relaxation time -- Aging mechanism -- State-of-Health estimation; Energy storage -- Periodicals; Energy storage -- Research -- Periodicals; 621.3126
• DOI: 10.1016/j.est.2022.105386
• Languages: English
• ISSNs: 2352-152X

Abstract

Abstract: Electrochemical Impedance Spectroscopy (EIS) is powerful tool to explain the degradation mechanisms estimate State-of-Health (SOH) of lithium-ion batteries. Recently, many online EIS measurement methods were proposed, which makes SOH estimations based on EIS data have more potential to achieve on-board applications. In this paper, the distribution of relaxation times (DRT) method is introduced to decouple the internal behaviors of lithium-ion batteries based on EIS data. And degradation mechanisms of all behaviors in both the cycling aging condition and the calendar aging condition are analyzed. In these analyses, similarities and differences of degradation mechanisms between these two aging conditions are discussed, which can provide the theoretical basis for developing battery degradation models. In SOH estimation section, correlation analyses between SOH and parameters extracted from DRT plots and EIS data are carried out and parameters showing linear relationship or piecewise linear relationship with SOH are selected from correlation analyses to achieve the highly accurate SOH estimation. Considering that State-of-Charge (SOC) of batteries may change when their EISs are measured online, we try to set parameters from different SOC as inputs of the SOH estimation. In this case, the DRT-based SOH estimation method still has an enough accuracy. In the end, the advantage in terms of computational efficiency and estimation accuracy and the adaptability for different regression methods of the DRT-based SOH estimation method are discussed. Highlights: The derivation and solving process of DRT are shown. Aging mechanisms of cycling aging and calendar aging are compared. Parameters showing piecewise linear trends are considered in SOH estimation. SOC changes in EIS measurements show little impact on accuracy of SOH estimation.